[0022] The present invention is a walking assistance system for visually impaired persons, which includes three parts: a head detection device, a waist detection device and a leg detection device, and the three parts communicate data through a wireless sensor network. The waist detection device is a network gateway, and the head and leg detection devices are network nodes. The head detection device mainly detects obstacles and approaching people or objects in four directions; the waist detection device detects left oblique front, front oblique front, right oblique front, left oblique rear, front rear, and right oblique rear six. The leg detection device is divided into 2 thigh obstacle avoidance devices and 2 calf obstacle avoidance devices, and each device can detect obstacles in three directions: front, back and outside of the leg, so that the whole system can The omni-directional detection of the surroundings of the human body is carried out, and the detection is completed by distance measuring sensors.
[0023] The head and leg detection device is mainly composed of four parts: vibration feedback unit, power management unit, ZigBee MCU control unit and infrared ranging unit; its feature is: ZigBee MCU control unit controls infrared ranging unit to transmit infrared signals and receive The infrared signal reflected by the obstacle analyzes whether there is an obstacle in the direction of the infrared ranging sensor. If there is an obstacle, the control unit transmits a warning signal for the visually impaired person through the vibration feedback unit. Through the use of ZigBee technology, distributed control of each detection device, no bus communication is required, and each detection device has its own independent power management unit, using a common reusable lithium battery as the power supply method, without the need for complicated power cords; Each detection device is tied to each part of the human body by a strap. In addition to the above four parts, the waist detection device also includes a stride frequency detection unit. The stride frequency detection unit is mainly composed of a three-axis acceleration sensor, which is used to detect the walking speed of the visually impaired during walking, according to the stride frequency and distance obstacle If there is an obstacle in the front, the feedback unit of the detection device will determine the vibration frequency of the vibration motor based on the step frequency of the visually impaired user, that is, the faster the speed, the shorter the distance, the higher the vibration frequency of the vibration motor.
[0024] The present invention will be further described in detail below in conjunction with the drawings and embodiments.
[0025] Such as figure 1 As shown, a wearable blind guide device according to an embodiment of the present invention includes: a head detection device 101, a waist device 106, a thigh device 102, 105, and a calf device 103, 104. Each device is worn on a person's body through a strap. Among them, the waist detection device 106 serves as the gateway of the wireless sensor network, and the other detection devices are network nodes, which form the network through self-organization. The head detection device 101 has obstacle detection units in the front, rear, left, and right directions of the wearer. The waist detection device 106 not only has obstacle detection units in three directions in front of the waist, but also has obstacle detection units in three directions behind the waist. The thigh devices 102 and 105 and the calf devices 103 and 104 are all facing forward. , Rear and lateral installation of obstacle detection unit. In this way, a three-dimensional detection circle can be established around the human body to effectively measure the distance with the infrared sensor. It can give warning information about people or things approaching to the obstacles around the visually impaired and to the approaching visually impaired.
[0026] Such as figure 2 As shown, it is a hardware block diagram of the head detection device in the blind guide system. The system mainly includes a strap 201 and four obstacle detection units 202, 203, 204, and 205. Among them, three obstacle detection units 202, 204, and 205 are mainly composed of vibration motors 210, 213, and 211 and infrared sensors 206, 208, and 209. As the communication unit of the head detection device, the obstacle detection unit 203 not only has a wireless communication function, but also has an obstacle detection function. It mainly includes power management 207, ZigBee MCU 212, infrared sensor 214 and vibration motor 215. The power management 207 of the obstacle detection unit 203 is the power management of the entire head detection device, while the ZigBee MCU controls the detection control of the entire head device, data communication, and performs feedback warnings.
[0027] The leg detection device is mainly divided into four sub-devices, left and right thighs and left and right calves, with the same structure. The hardware block diagram of the detection device is as follows image 3 As shown, it mainly includes a strap 301 and three obstacle detection units 302, 303, and 304. The obstacle detection units 302 and 304 respectively include infrared sensors 305 and 307 and vibration motors 308 and 312, and the obstacle detection unit 303 is mainly composed of a power management 306, a ZigBee MCU 309, an infrared sensor 310 and a vibration motor 311. The power management 306 and the ZigBee MCU 309 have the same functions as the corresponding parts of the head detection unit.
[0028] Such as Figure 4 As shown, it is a hardware block diagram of the waist detection device, which mainly includes a strap 401 and 6 obstacle detection units 402, 403, 404, 405, 406, and 407. The obstacle detection units 402, 403, 405, 406, and 407 have the same structure, and they are all composed of an infrared sensor and a vibration motor. The obstacle detection unit 404 mainly includes five parts, which are power management 409, ZigBee MCU 417, three-axis acceleration sensor 416, infrared sensor 421, and vibration motor 422, respectively. Different from the head and leg detection devices, the ZigBee MCU 417 of the waist detection device has to undertake the task of constructing a wireless sensor network, as well as analyzing and fusing data information sent by other detection devices, and feeding back the analysis results. At the same time, the waist detection device also includes a step frequency detection function, which mainly collects data from the three-axis acceleration sensor 416 to obtain the walking speed of the person.
[0029] Such as Figure 5 Shown is the program flow chart of the entire system. After the system is initialized, the data is collected by the detection devices at the four positions of the head, waist, thighs and calves, and transmitted to the gateway in a fixed period. The gateway performs comprehensive data analysis and obtains the sensor data from each detection device. , Can extract road pits, hollow obstacles, suspended obstacles and ground raised obstacles information. Know the walking speed of the blind from the acceleration sensor data, and give warnings according to the type of obstacle and the distance of the obstacle, the type of obstacle can be learned from the position of the vibration, and the distance of the obstacle according to the frequency of the vibration. The specific data analysis process such as Figure 7 , 8 , As shown in 9.
[0030] Such as Image 6 As shown, it is an example of a device user wearing, and each detection device is worn as shown in the figure. The drawings show the left and right views of the device after being worn by the user. The figure on the right shows the front and back views of the device after being worn by the user.
[0031] Such as Figure 7 As shown, the visually impaired person detects various obstacles in a standing state after being worn. The detection devices of the entire system are distributed with a substantially fixed height difference from the head to the calf, so that the entire system has a better detection of obstacles of different heights. As shown in the figure, there are four physical obstacles standing on the ground, and obstacles of different heights are detected by different numbers of sensors. The detected data is sent to the gateway to analyze the obstacle height judgment, and the vibration motors in all detection devices covered by the obstacle work and give warning information. When there are only obstacles detected by the calf detection device, the vibration motor in the calf detection device will give a warning. The vibration frequency of the vibration motor can be adjusted according to the walking speed and obstacle distance. The higher the vibration frequency, the more the obstacle near. When both the calf and thigh detection devices detect obstacles, the vibration motors of the two parts give warnings. By analogy, obstacles protruding from the ground can be detected and effective warning information will be given. Of course, for the hanging obstacles, the visually impaired person on the surface suffers head injuries. When only the head detection device detects the obstacle, the head detection device gives a warning about the hanging obstacle, allowing the visually impaired person to lower his head and pass forward. For obstacles that are hollow like tables, they can only be detected by discrete detection devices. For example, the thigh, calf, and head detection devices are not detected, but only the waist detection device detects it, indicating that the front is a hollow Obstacles, the visually impaired people still need to be prompted to detour.
[0032] Such as Picture 8 Shown is a step-by-step diagram of a person walking normally. In the process of walking, the relative positions of the devices worn on the thighs and calves change continuously over time, but the overall change is cyclical from 1 to 4. The detection directions of the infrared sensors installed in the detection device are all horizontal, and the sensor direction of the calf changes into a periodic process of "up-flat-down-down". In this way, there is an obstacle in front that can be detected by the infrared sensor of the calf, and the graph made by the distance data output by the infrared sensor is a form of damped vibration. The information on whether there is an obstacle and the distance of the obstacle is extracted from the characteristics of the curve, and the vibration frequency of the button vibration motor 201 is fed back to the device user. Similarly, the data returned by the thigh device is also processed by the above method.
[0033] Such as Picture 9 Shown is a schematic diagram of the system's detection of ground pits. Since the relative positions of the thigh and calf alternate during human movement, it is usually only to extract the distance information of the infrared sensor from the obstacle when the leg is straightened in the natural state. In the present invention, the infrared distance data of the leg is fully utilized to extract the infrared distance information of the calf in the unnatural state, so as to obtain the basic road condition of the road surface and whether there are pits. As shown in the figure, when a person walks, there are four steps from 1-4. In the first step, the left calf can detect the ground distance DL(1, i), and the second step, the left calf can detect the ground distance DL(2, i). ), the third and fourth steps are divided into DR(1,i) and DR(2,i), where i is the i-th step in the walking process. When the ground is relatively flat, DR(j,i) and DL(j,i) remain basically unchanged, and when a pit is encountered in the figure, the data will inevitably be distorted, so that timely warnings can be made The information is fed back to the visually impaired to avoid falling.
